A genetic algorithm optimized ANN-based MPPT algorithm for a stand-alone PV system with induction motor drive

This paper describes FPGA implementation of a Maximum Power Point Tracking (MPPT) for Photovoltaic (PV) applications. By slightly modifying the original algorithm, an improved variable step-size P&O algorithm is realized and efficiently implemented using a hard-ware description language (VHDL). Subsequently, the new MPPT algorithm integrated with a solar-powered battery charging system is implemented on the XC2C384 FPGA without external sensor unit requirement. Experimental results with a commercial PV array show that the proposed algorithm outperforms the conventional controller in terms of tracking speed and mitigation of fluctuation output power in steady state operation. The overall system efficiency is well above 96%.

FPGA Implementation of MPPT Using Variable Step-Size P&O Algorithm for PV Applications

Optimal battery charging, Part I: Minimizing time-to-charge, energy loss, and temperature rise for OCV-resistance battery model

This paper presents a new functionality for high-power battery chargers by incorporating an impedance measurement algorithm. The measurement of battery impedance can be performed by the battery charger to provide an accurate equivalent model for battery management purposes. In this paper, an extended control capability of the onboard battery charger for electric vehicles is used to measure the online impedance of the battery. The impedance of the battery is measured by the following: 1) injecting ac current ripple on top of the dc charging current; 2) transforming voltage and current signals using a virtual $\alpha{-}\beta$ stationary coordinate system, a $d{-}q$ rotating coordinate system, and two filtering systems; 3) calculating ripple voltage and current values; and 4) calculating the angle and magnitude of the impedance. The contributions of this paper are the use of the $d{-}q$ transformation to attain the battery impedance, theta, and its ripple power, as well as providing a controller design procedure which has impedance measurement capability. The online impedance information can be utilized for diverse applications such as the following: 1) a theta control for sinusoidal current charging; 2) the quantifying of reactive current and voltage; 3) ascertaining the state of charge; 4) determining the state of health; and 5) finding the optimized charging current. Therefore, the benefit of this method is that it can be deployed in already existing high-power chargers regardless of battery chemistry. Validations of the proposed approach were made by comparing measurement values by using a battery charger and a commercial frequency response analyzer.

Online Embedded Impedance Measurement Using High-Power Battery Charger

This paper presents a sinusoidal ripple current charging algorithm based on embedded impedance measurements. Existing battery charging strategies typically do not take into account the electrochemical properties of batteries, because these factors are difficult to obtain during charging operation. Factors of concern include lithium plating, growth of a solid electrolyte interphase, limited exchange current, and slow diffusion rates. It is beneficial to utilize these parameters during charging operation, because the charging current can adapt to the time-varying characteristics of a battery. Consequently, battery life cycle, charging speed, and charging efficiency all improve. In this paper, rigorous analysis of electrochemical characteristics is performed and a method for minimization of variations of charge transfer impedance is explained based on a sinusoidal ripple current charging algorithm. To obtain the optimal ripple current frequency, ac impedance analysis based on the dq transformation method is proposed. As a result, this method improved charging efficiency and reduced lithium plating by activation polarization. Simulation and experimental results using a 14.6-V LiFeMgPO4 battery are used to validate and demonstrate the performance of the proposed control scheme. Based on the proposed control scheme, the charging time and efficiency of the Li-ion battery are improved by 5.1% and 5.6%, respectively.

Electrochemical State-Based Sinusoidal Ripple Current Charging Control

This paper presents the development of a maximum power point tracking algorithm using an artificial neural network for a solar power system. By applying a three layers neural network and some simple activation functions, the maximum power point of a solar array can be efficiently tracked. The tracking algorithm integrated with a solar-powered battery charging system has been successfully implemented on a low-cost PIC16F876 RISC-microcontroller without external sensor unit requirement. The experimental results with a commercial solar array show that the proposed algorithm outperforms the conventional controller in terms of tracking speed and mitigation of fluctuation output power in steady state operation. The overall system efficiency is well above 90%.

A Solar-powered Battery Charger with Neural Network Maximum Power Point Tracking Implemented on a Low-Cost PIC-microcontroller

This paper presents the development of a maximum power point tracking algorithm using an artificial neural network for a solar power system By applying a three layers neural network and some simple activation functions, the maximum power point of a solar array can be efficiently tracked The tracking algorithm integrated with a solar-powered battery charging system has been successfully implemented on a low-cost PIC16F876 RISC-microcontroller without external sensor unit requirement The experimental results with a commercial solar array show that the proposed algorithm outperforms the conventional controller in terms of tracking speed and mitigation of fluctuation output power in steady state operation The overall system efficiency is well above 91%

This paper presents the development of an intelligent genetic algorithm (GA) technique for training of a generalized regression neural network (GRNN) controller to achieve a compact network and to decrease battery charging time on a cost-effective RISC microcontroller. The suitable input-output data were selected from GA mechanism to establish GRNN. The computational complexity of GRNN can be reduced replaced by some simple polynomial forms. As a consequence, the fast charging device for Nickel-Cadmium (Ni-Cd) batteries can be efficiently implemented on a low-cost 16F876A RISC microcontroller. Experimental results are shown to demonstrate superiority of the proposed system.

Implementation of GA-trained GRNN for Intelligent Fast Charger for Ni-Cd Batteries

This paper presents an intelligent technique for training the neural network controller in order to archive a compact network and to decrease battery charging time. An ultra fast charging device for nickel-cadmium (Ni-Cd) batteries is designed through the generalized regression neural network (GRNN) and implemented with the MATLAB/SIMULINK for testing and operating on real system. The input-output data for training neural networks were collected from rigorous experimentation. The suitable data were selected to establish GRNN comprising only 13 processing elements. Each node of the RBFs is an extendable support function which recovers the drawback of the existing compact support radial basis functions (CSRBF). Experiments with real time implementation clearly show that the proposed technique not only requires less neural processing units but also yields less MSE than the RBF technique.

Intelligent ultra fast charger for Ni-Cd batteries

This paper presents an intelligent genetic algorithm (GA) technique for training of a neural network controller to achieve a compact network and to decrease battery charging time An ultra fast charging device for Nickel-Cadmium (Ni-Cd) batteries is designed through the generalized regression neural network (GRNN) and implemented with the MATLAB/SIMULINK for testing and operating on real system The input-output data for training neural networks were collected from GA The suitable data were selected to establish GRNN comprising only 13 processing elements Each node of RBFs is an extendable support function that overcomes the drawback in the existing compact support radial basis functions (CSRBF) Experiments with real time implementation clearly show that the proposed technique not only requires less neural processing units but also yields less MSE than RBF technique

W. Kiranon

Papers

A Solar-powered Battery Charger with Neural Network Maximum Power Point Tracking Implemented on a Low-Cost PIC-microcontroller

A Solar-powered Battery Charger with Neural Network Maximum Power Point Tracking Implemented on a Low-Cost PIC-microcontroller

Implementation of GA-trained GRNN for Intelligent Fast Charger for Ni-Cd Batteries

Intelligent ultra fast charger for Ni-Cd batteries

GA-trained GRNN for Intelligent Ultra Fast Charger for Ni-Cd Batteries